It is probably wise to begin with the epidemiological data, since these are easiest to verify and likely to be the most reliable. Three incontrovertible statements can be made: ulcerative colitis and Crohn’s disease tend to run in families;’both disorders are associated with an increased risk of colon cancer; 2, 3 and ulcerative colitis is commoner in non-smokers or ex-smokers than in people who smoke, whereas Crohn’s disease is commoner in smokers. 4, 5 Although ulcerative colitis and Crohn’s disease tend to run true to type in most families, there are many examples of families with cases of both disorders. We should, therefore, not discount the possibility that there is a common inherited factor predisposing to inflammatory bowel disease, and that environmental factors determine the phenotype-ulcerative colitis or Crohn’s disease. For example, Hodgson’suggested that individuals who inherit susceptibility to inflammatory bowel disease might be more likely to develop Crohn’s disease if they smoke and ulcerative colitis if they do not. The nature of the link between inflammatory bowel disease and colon cancer also needs careful consideration. Although it is commonly assumed that colitis causes the cancer, extensive epidemiological data from a three-centre study show that the relative risk of colon cancer in patients with ulcerative colitis compared with age-matched controls is constant irrespective of disease duration. This finding could suggest co-inherited risks for inflammatory bowel disease and cancer rather than a cause-and-effect relation between inflammatory bowel disease and cancer. The lack of any significant increase in rectal cancer in patients with localised ulcerative proctitis? suggests that chronic inflammation itself is unlikely to be the cause of increased cancer risk; this observation could be interpreted as showing that such patients have a minor inherited abnormality that results in less extensive colitis and a correspondingly slight increase in risk of cancer. The greater relative risk of cancer in patients who develop colitis at an early age 1, 7 similarly implies a greater degree of genetic defect rather than a cause-and-effect relation between colitis and cancer; again this idea is supported by a constant, albeit much greater, relative risk compared with age-matched controls. 3 Inflammatory bowel disease and related colon cancer could, therefore, all be linked to the same inherited factor. If so, there ought to be an increased risk of colorectal cancer in relatives of patients with inflammatory bowel disease; no studies have yet addressed this possibility.